[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Public WebGL] question about implementation of anti-aliasing; and a remark for benchmark authors



Hi,

Two parts in this email:

*** Question about implementing anti-aliasing ***

In an earlier thread, Adrienne kindly pointed me to Chromium's implementation of antialiasing:
http://src.chromium.org/viewvc/chrome/trunk/src/chrome/renderer/webgraphicscontext3d_command_buffer_impl.cc?annotate=70474#l79

But in Mozilla's implementation, we do all the rendering into a FBO.

My understanding was that just requiring samples and sample buffers at GL context creation time was not enough to get multisampling on FBOs, and that for that I needed the GL_EXT_framebuffer_multisample extension (part of OpenGL 3.0).

Is that correct?

If yes, how does Chromium/WebKit handle this problem? Are you also doing your rendering into a FBO, or a PBuffer?

*** Remark for benchmark authors, and Firefox status ***

Firefox doesn't currently do anti-aliasing. Chrome does. The 'antialias' context attribute is 'true' by default. This means that by default, Chrome does higher-quality rendering.

So benchmarks trying to compare browser performance should consider using explictly antialias:false so as to avoid putting Firefox at an unfair advantage.

I probably won't be able to get antialiasing accepted into Firefox 4.0 at this point, given the risks in terms of exposing more unforeseen driver bugs at the last minute. Probably for the next version.

Benoit
-----------------------------------------------------------
You are currently subscribed to public_webgl@khronos.org.
To unsubscribe, send an email to majordomo@khronos.org with
the following command in the body of your email:
unsubscribe public_webgl
-----------------------------------------------------------